Tos FN22 milling machine (Deckel clone)

[Agroshield]

If it wasn't so damned difficult to move I'd... ...plane the front flat.

If you cannot take the casting to the planer, can you take the cutting tool to the casting and machine it in-situ?

If you fit the Z- and X-axes, you have something that moves up, down, left and right under precise control to which you could mount a cutting tool of some sort.

The big issue would be that it would not cover the whole surface in one setup. However, if whatever it is could be rotated 180 degrees, it could be made to do so.

Imagine a cutting tool in the spindle of the dividing head below (with the overarm and outboard support removed). With the two rotational adjustments on the angle plate, you can point the sharp end of the tool at every point of the surface you want to machine.

If you can find a way to mount it, the powered spindle below would do. Where it attaches to the Y-axis is a reference surface and if that surface is fixed to your X-axis, it is rigid enough and long enough to be able to overhang and use a flycutter.

 

[Pete.]

I do appreciate the idea but if I got that desperate I would rather make the effort to get it over to Clive for planing. I have a few steps I need to take before I even get that far. On a mill, there is a lot more to consider by way of alignment than on a lathe because the part moves, whereas on a lathe it just rotates.

 

[awemawson]

Pete you are obviously working to very fine measurements here. When the casting resumes it's normal attitude and stands up, will it's own weight not perhaps distort it by amounts of similar magnitude. And again when the table / saddle arrangement gets hung on the ways there must be some. albeit small distortion.

 

[Pete.]

Pete you are obviously working to very fine measurements here. When the casting resumes it's normal attitude and stands up, will it's own weight not perhaps distort it by amounts of similar magnitude. And again when the table / saddle arrangement gets hung on the ways there must be some. albeit small distortion.

I thought about that Andrew and referred to it in post#344. I don't think so as long as I have it resting back on it's frame. Truth is, there's no way of knowing until I get the work done and bolt it back on the base. When I do that I'll measure it again and if it's shifted I'll have a little cry then set about shimming the base or re-scraping. Right now, I'm fairly confident that the casting is stiff enough to resist it.

Putting the console table and then the work tables on is another thing entirely. I'm certain that it WILL droop but I have no experience to draw upon to gauge how much. All I can do for now is make dame sure the Z and Y ways are perpendicular to each other because they control the direction of travel. A drooping table can be rectified by scraping back the top way of the console table or x-slide way, which is easy to do later on in the work.

 

[Martin-Sy]

Awesome @Pete.
When are you running your next machine tool rebuild course

 

[Pete.]

Awesome @Pete.
When are you running your next machine tool rebuild course

Not me I'm just an enthusiastic amateur who's showing what he's up to

 

[Pete.]

Right then bit of a long post you'll have to bear with me whilstI try to run through this all in a way that makes sense.

Tonight I set about checking the long ways I just scraped to see if they are running parallel to each other. To do this I set my very best iron surface plate on 3 jacks above the pair of ways.



Then I took my surface gauge with tenth-reading tesa DTI and used that to set the plate up. Firstly I set both of the jacks on the right so that I had the same reading on the DTI. I had to make sure the DTI was directly in line with both jacks when doing this or else the readings would be meaningless. After that I set the DTI on one side at the other end and adjusted the single jack until that also read zero with the first pair.

This is running the DTI down the side I levelled.





So, if the ways are parallel then the other side should also read zero-zero along the 14" length





Clearly, they don't, this RH way is running out from the LH one by 0.0007" over 14". Bit annoying but I didn't expect it to be right.



Now, usually what do to combat this on something like a cross-slide I scrape the cross-slide flat ways all together on a surface plate so that I know they are all on one plane then I use that to spot the saddle ways to check if they are parallel just like I'm doing here. I'd use a straight egde to actually scrape the ways in and the cross-slide to check I'm not running out of parallel. It works great. I can't do that here because 1. the console slide is too heavy for one person to do that and 2. it has a big lump hanging off one side so also out of balance. Hence the checking I'm doing here.

 

[Pete.]

Now, whilst the above was going on I had not only the plate sitting on 3 points but the casting also sitting on it's frame, the load bearing on 3 points also. This is done to try to prevent any twist in the frame cause by having 4 or more points of contact. Even so I was still very wary of the point @awemawson made that the casting might be distorting because of how it's being held.

I also had to think about the relationship of these long ways to the y-axis ways that are also already scraped. They have to be 100% perpendicular to each other because these ways control the direction of travel of the part being machined. (One point to note these ways do not have any influence on table droop). So I thought I would kill 2 birds with one stone by putting the casting back to it's tilt-angle, re-checking the relationship of the two ways to each other and in the process check their relationship to the y-axis ways (which I know are 100% parallel to each other because I scraped them together using a granite plate).

Here's my setup. I got the casting back down from horizontal but still on 3 points to start with to allow me to set up my granite square without needing 5 hands. I set a piece of scrap cast iron across as a ledge to hold the granite, squared across the ways using my box square. Then I put the box square across the ways to hold the granite exactly perpendicular and the granite rested against it and against the way and I let gravity hold it in place.



The indicator I put on a small mag base on my @pressbrake1 18" straight edge which would give me a nice smooth movement for measurement. This is a pic of the opposite side being measured with the same setup.



So I ran the DTI up and down the granite square and took note of the readings. This is the worst side.





It shows a drop of ten and a half tenths over 6". (I know I wrote 11 but the dial says 10.5). The opposite side shows a drop of 0.0007" over the same distance.



So what does this mean? Well, it's not too bad actually. If the ways were actually parallel I would call it done. The ways would be square to 7 tenths in 6" which is just under thou and a half in a foot. I could live with that quite happily. I doubt very much that I will match that with the further relationships. However the ways are not parallel so if I am going to fix that I might as well fix both.

More importantly it tells me that my housing is not getting twisted about any appreciable amount by me changing the angle on the frame. The difference in these two measurements is 0.00035" in 6" which is 0.0008" in 14". That is only 0.0001" different from my original measurement using the surface plate up on the jacks.

As a final check I didn't know if I could trust my Chinese granite sqaure so I re-measured one side using the box square.



Got the same measurement exactly so I'm satisfied with that too.

 

[Pete.]

So what happens next? Firstly , it hurts me to do it but I'm going to have to re-scrape the flat way that runs down the most. It's 30" long so I'm going to have to bring one end down exactly a thou and a half. That should get them parallel but still running 'down' towards the y axis making the angle between slightly greater than 90 (by hopefully 0.0007" in 6" each). It would be stupid of me to try to rectify THAT by continuing to scrape a pair of 30" ways when I can achieve the same by scraping the much smaller y-axis ways down at the back so that is what I will do. In truth, that is probably the source of the error anyway if you recall I mentioned that the y-axis way was worn at the front from all the overhanging weight and I was already biasing the work more to the rear to compensate. I just didn't do it enough.

 

[Agroshield]

I may have missed something, but how are you ensuring the ways are 'level' with each other?

In the first picture:

in the axis of the pair of jacks, the surface plate is the same distance above the ways. But what if the foreground one is 5mm closer to the garage floor than the background one? If it is, it will skew the X- axis relative to the Y- when you look at in plan view. Your test with the granite square is ensuring things are correctly square in side view, but there is an additional need for squareness at right angles to this.

I do not know if it will help, but consider the garage floor as a flat, level plane. For the Z- and Y- ways, ask what needs to be plumb to, parallel with and the same distance above that plane. You need to ask theat question in two mutually perpendicular directions.

 

[daedalusminos]

How do you ensure the ways are not tilted across their width?

 

[Pete.]

How do you ensure the ways are not tilted across their width?

Well it doesn't actually matter so long as the mating part matches for contact. It's no different to having a lathe bed with flat ways or vee ways. Of course, the original ways would have been machined in a single setup so they would be parallel and co-planar then on a machine where they were scraped the scraping hand wwould have used a jig most likely. This goes back to my previous post about doing cross-slides. I scrape the cross slide bottom ways on a surface plate and they make an excellent gauge. Can't do this here.

The only options are to either just scrape it and adjust the console table ways to suit. Not a great option TBH because 1. the console ways are already scraped flat, on a surface plate and 2. it's a lot harder iron and it has oiling grooves in it. Another option is to get someone to help me manipulate the slide onto the ways to check them then just adjust as required. Another option would be to make a short gauge which has the sole purpose of checking these for co-planar. That would be easy enough if I had a working milling machine

I may have missed something, but how are you ensuring the ways are 'level' with each other?

In the first picture:

in the axis of the pair of jacks, the surface plate is the same distance above the ways. But what if the foreground one is 5mm closer to the garage floor than the background one? If it is, it will skew the X- axis relative to the Y- when you look at in plan view. Your test with the granite square is ensuring things are correctly square in side view, but there is an additional need for squareness at right angles to this.

I do not know if it will help, but consider the garage floor as a flat, level plane. For the Z- and Y- ways, ask what needs to be plumb to, parallel with and the same distance above that plane. You need to ask theat question in two mutually perpendicular directions.

Ok I see what you mean. If you presume that the y-axis is perfectly vertical (with the machine lying on it's back) then how do you know that the line drawn through the ways I'm checking is perfectly horizontal? Well, the ways probably are close but likely not bang on. It doesn't matter though so long as it's close because the x-axis to y-axis relationship is referenced from the y-axis angled ways which I haven't touched yet. They will certainly have wear, will certainly need scraping. I just have to make sure I scrape them in the right direction to improve the error not make it worse.

 

[minimutly]

You're into a whole world of pain here if you cock this up (which I don't expect for a minute you will). But, you need to realise that the film of oil you want between the ways will be some way thicker than the error you might be correcting?
As for checking co-planar of ways I used a pair of 123 blocks and a ground edge across the top (to get above the dovetails) at increments along.

 

[awemawson]

Milling machine here Pete at your disposal

 

[Windy Miller]

Milling machine here Pete at your disposal

Likewise here @Pete.

 

[Pete.]

You're into a whole world of pain here if you cock this up (which I don't expect for a minute you will). But, you need to realise that the film of oil you want between the ways will be some way thicker than the error you might be correcting?
As for checking co-planar of ways I used a pair of 123 blocks and a ground edge across the top (to get above the dovetails) at increments along.

I can do the same by re-setting the surface plate a little higher up and using the surface gauge inverted. It's a comparative measurement not an absolute one so any error will be immediately obvious. After I set the thing up last night it occurred to me that I should have raised it even more then I would have been able to work across as well as along.

I'm not worried about messing it up if I did nothing further at all it would still be a very nice machine.

I don't worry about oil film thickness because it applies to all parts of the way equally.

Another thing that taking time to straighten these two ways up will do is create two references that I can use to scrape in the flat section in between that is currently all over the place. That will come in very handy when setting up the x-y-z relationships as I can put a granite square against it to check the y axis and a dial on the console table to check the x.

 

[Pete.]

Milling machine here Pete at your disposal

Likewise here @Pete.

Thank you both I appreciate it.

 

[Agroshield]

I just have to make sure I scrape them in the right direction to improve the error not make it worse.

That is the crux of it. If you scrape X off something to make it correct and then have to scrape 2X off another part to bring it to its correct alignment, it is making work for yourself.

I wonder if anyone has ever made a comparison between the ease of aligning a Euro-style mill and a traditional knee mill.

 

[Pete.]

That is the crux of it. If you scrape X off something to make it correct and then have to scrape 2X off another part to bring it to its correct alignment, it is making work for yourself.

Yeah, that's what happens. I'm still very much a novice so if I had more experience I would have done these checks with it rough-scraped in then adjusted it out doing the next round. Of course, a professional rebuild would have had it machined/planed/ground before scraping so the geometry would be mostly (if not completely) there. It might only have needed machining then flaking in fact. Here in my home shop I have no choice but to do all the major and minor corrections by hand.

I wonder if anyone has ever made a comparison between the ease of aligning a Euro-style mill and a traditional knee mill.

Don't suppose there's a lot in it. From what I can see the euro style mills tend to have bigger ways for the work envelope so there's more physical work but the knee ones are IMO more tricky to align.

 

[Pete.]

So this evening I've set about correcting the small amount of out-of-parallel of the main flat ways. The ways are 30" long so i separated them into 10 3" sections and numbered them starting from zero to nine. (Zero because the first section doesn't get scraped, then each next section gets scraped one round more than the previous). The idea is that by scraping these (very shallow) steps you can turn the whole face as much as you need.

Before scraping I set the plate back up to measure how out of flat (relative to each other) the ways had got across the way. One side was a few tenths out of flat the other side only about a tenth and a half. This is the better side. I will correct the other side later if required, I want to get this one straightened out first.





So I set about scraping from 1 to 9. I started off doing No1 once, No2 twice, No3 three times like this.



Then figured I would have a large step to deal with when I got to 9 so I changed tack and did 4-9 four times then scraped 5-9 then 6-9 then 7-9, 8-9 and finally 9 on it's own. Also, after I did 3 rounds I realised that it was going to be very easy for me to lose track and risk doubling-up so re-did the numbering and used ticks to record each time I scraped.



Big difference between roughing and finish scraping, and it shows up in the initial print.





I did a few rounds of scraping just to make the surface somewhat sensible to get a reading from with the surface gauge and DTI.



Then I set the plate back up so it was dead parallel the other side and took a reading from the side I'd just scraped. Over the 16" plate it was showing just under 5 tenths of rise still. This is better than before but not nearly as good as I was expecting. I was actually worried that I might have over-shot.





Then I realised I had set the plate up wrong. I should have set it near the zero end of the way so that my initial reading was 'at zero' but what I actually did was set it at the same spot as before. This put the back of the surface gauge in No1 and the front in No2 but there's a bit of a dip at 2-3 which I hadn't scraped out yet (it's hard to seebecause the the LED lights reflecting in that spot) so the gauge was sitting a tenth or two 'nose down' for my initial reading. I don't know why it was deeper there I can only put it down to scraping deeper at the start of the session and less deep as my arms started getting tired.

Anyway, I'm on the right track, I'll give it a few more rounds tomorrow to scrape out the dip at 2-3 and bring the bearing in more then I expect that it will be as close as it's ever going to be to perfect.